Exporting data from OSM#

OpenStreetMap is a collaborative, open-source mapping project that provides free, editable geographic data for the entire world. It’s often called the “Wikipedia of maps” because anyone can contribute — and millions of users do.

The data includes a wide range of real-world features such as:

  • Roads, buildings, parks, rivers

  • Public transport stops, bike lanes, and walking paths

  • Points of interest like schools, hospitals, cafes, and more

Thanks to its openness, flexibility, and global coverage, OSM has become one of the most widely used spatial data sources — powering applications in urban planning, transportation, humanitarian response, and research. (You can fin more Open Data Sources here)

And the most of the analsysis in this course will be based on this data

OpenStreetMap data can be exported in many ways — through web interfaces, GIS software, or APIs. But since we’re working in Python, we can fetch OSM data directly in our Jupyter Notebook, which makes the workflow faster and fully reproducible.

To do this, we’ll use a powerful library called OSMnx

Import libraries#

import osmnx as ox

  • OSMnx (osmnx) — a Python library for downloading, analyzing, and visualizing OpenStreetMap data. It allows you to easily retrieve geospatial features such as buildings, street networks, and boundaries by place name or coordinates. OSMnx integrates smoothly with GeoPandas and is ideal for urban analysis, network science, and working with real-world geographic data — all directly within your Python environment.

Defining the area#

First, we need to define the area for which we want to get the data. Let’s choose the Central District of Saint Petersburg. We’ll use the geocode_to_gdf function to retrieve the district boundaries.

admin_district = ox.geocode_to_gdf('Wieden, Vienna')

Let’s see what we’ve got. In the explore method, we’ll set the tiles parameter to change the default OSM basemap to Positron.

admin_district.explore(tiles='cartodbpositron')
Make this Notebook Trusted to load map: File -> Trust Notebook

Incredible! But this is exactly what we need — the boundaries of the Central District of Saint Petersburg.

Exporting features#

Now let’s try exporting building data for the Central District. To do this, we’ll create a tags variable and specify the key for the objects we need — buildings. We’ll then pass the district name and the tags to the geometries_from_place function.

You can read more about OSM keys and values here.

tags = {'building': True}   

buildings = ox.features_from_place('Wieden, Vienna', tags)

buildings.explore(tiles='cartodbpositron',tooltip=None)
Make this Notebook Trusted to load map: File -> Trust Notebook

We’ve got the buildings!

Now let’s filter the data to include only Polygon and MultiPolygon geometries.

buildings = buildings.loc[buildings.geom_type.isin(['Polygon', 'MultiPolygon'])]

And now we can safely save the data!

Spoiler — an error will occur during saving because some fields contain data in list format. This format isn’t supported for saving, so we’ll fix it by converting those fields to strings.

This loop checks each column in the buildings GeoDataFrame to see if it contains list-type values. Some columns in OSM data (especially tags like amenity, building:use, etc.) can contain lists — for example, when multiple values are assigned to one feature.

However, file formats like GeoJSON or Shapefile do not support list values, which can cause errors when saving the data.

for col in buildings.columns:
    if any(isinstance(val, list) for val in buildings[col]):
        buildings = buildings.astype({col: str})

  • It loops through each column in buildings.

  • For each column, it checks if any of the values in that column are lists.

  • If so, it converts the entire column to strings using astype({col: str}), so that lists become plain text (e.g., "['shop', 'office']").

Saving to file#

#buildings.to_file('../data/osm_build.geojson')

This command saves the buildings GeoDataFrame to a GeoJSON file, which can then be opened in any GIS software or reused in other projects. The file format is automatically determined by the file extension (.geojson in this case).

Summary#

In this section, we explored how to access and use OpenStreetMap (OSM) data directly in Python using the OSMnx library.

We learned how to:

  • Define an area of interest by name and retrieve its boundaries

  • Download OSM data such as buildings or street networks

  • Convert OSM data into GeoDataFrames for further spatial analysis

  • Visualize the results on an interactive map — without leaving the notebook

Using OSMnx, we can quickly integrate real-world, up-to-date geographic data into our analysis workflows — making OSM a powerful and accessible data source for spatial projects.